'; .UNKAGESBETWEENT-GROUP - _ . g _ TRAlNER PERSONALITY VARIABLES _ ‘ AND TRAINER EFFECTIVENESS , ‘ o . - 0 . . . . o I. . .‘g ‘ . . . ~_'_, ‘ . .‘. . ... . '. ‘ ' . .-. ' ‘>‘ . .- - .- ‘ . A f u _ , _ .- u . . . n I a - . -.b ’ . - I' — .|:-. . 0- . . n ‘, - V . q A ' ' ' , . r . . - _ . ‘. . . ’ 1 u ' * ' ° ‘. 0 . - . ‘ - o - ~ , n - . . A . o , _ I ‘ . _- ~ A C . . u . _ . - . , . - . . . ..- . ‘ . . _ - . I - «. . . . . . _‘ ‘ ' . . _ , . , o~ ‘ V ' . . . Q n _ ‘ .3' ' ~ , . ‘ - ' ' ', ' 9 r ' - i v ‘ . I . , _ . ‘i O . u - ~ . o - . . . , _ I I . — n o . . - . . o . o . - ’ . . . . ‘ a A . - .. - II . . . . - ‘ - ‘ - . . - ¢ ‘ i - . v o o - _ .‘ — 7 . ‘ I O V . . u l . . . ‘ _ . _ . .- - ' ‘ . I r O . fl ‘ - - 4 g u ’ ‘ — A ‘- . . . .. . . . . . _ .— o . o _ . ‘ .'T . ' , . . . . ’ . g _ . h ' . - . - . T I a. ' ' . - -' . u . ~- , . Z . ‘ ' ’ O ' - . ' l— . c t . ' - ' ' . o . _ ‘ . , - . ’ . . _ ’ — ' ‘ ‘ V‘ o o l . . - ' ' 1' _ ‘ p . . - " ' . . " - - -'o . . . .| O ‘ - O - 1‘ ‘—- ‘ v . ' I .' -‘ , .' a a.» lo. . I . I > t O I ‘ — . 'u. ‘0 -» ' I . . 1 , ‘ - z ' '1‘ - l . . . ' , . I. a ’ .'. . ' P - o — ‘ . . . . . - , _ i . . ' ' . . . .- , — - . y . . . . -.- o . - y l .¢ . a V’_o ,' o . ’ _ ... r v , .4 ‘ - . _ _ ’ . . .1 p , - u . ' _ . . . ' . - ’ ' . g. ., , - . . . ' I I ‘ . ’ 17: . - . . . - ' . » . u A - 4 ' 5 - I ‘ F . t / ' ~ , I ' . 4 . . a . ' . . , a - _ l , - . , .-- 1 ‘ l ' ‘ ' I - . . l- a ‘ . - . '0 -‘ ’ I . o O c y . - — o . 0‘ .. - o . ‘ . n r o o I ' a' O . . ' - K 0- . ,,- - . . . ... . > . . ¢'n f a 2- . . ‘ ',' ‘,‘:, .:- o' . , I . _ ' ’g- . . . ‘ . - o o o I '. . .‘ . . , o . ' . - - o ‘ .' ' , . . r . ,._ o . " . - - .- J . - — a ‘ . ‘ -. ' - ' " ‘ ' :o"' -'- -4 . ’--' - ‘ . . . . . A .‘ - — - . - . - ‘ . _ . a . - -'_ , ' _ ~ ‘ ‘ o ‘ 0 . V ' ' ‘ ' , , — . ~ ‘. a , . 0‘ , . , v , . . . o ' , l' ' — , , .... ‘~ - - r . .4. . v o —. - - a_ . ' - ‘ . "V . . I _ . ' o I ' .. . .'.‘. ' .|,v ‘_- o , - a ‘_I~' :. ‘ . A v ' . . . I .v.' 1‘; “. . .I_ - u I ‘ ' ' . p -‘ - - l ‘. l l ‘ ' . ' 'v — p, l - .— . -. D_-'- C ‘ ‘ ‘ ' . '. , 'I - - v‘.‘ '-~ ', ' ‘ . ' ' ’ ‘ ~. .. I - ‘ I ' ' ’ . . o ,- .‘o‘, - ' .' - . o- - u ‘ ’ C ' v 0.- . ’ ' . ,- . ' " I .. .- o o — . . ,‘ ' ° ' ‘ . , ' _ m . , a ‘ ‘ ' ' ' ‘- 1' a l J y . .I .‘ v . r '. I, . . . . '.. ' '. ..§ ' .O . ... .' C I I v o ' ' ‘ .‘ . , - . . a d . a , . ' 4 ; l . ..‘ *I‘ _ - . ‘ .‘. - 4', . ' A .. ‘ A ’ IA ‘ . ’ ' - .’ I ' ' , . .90. ' ' _ '. . 1 .’ , . a ' -. '0’ (v P n ’ fl' '_ - ‘ n . ' I '- l' r y . ‘ . l r-‘ ‘ ‘ o ‘ v. '- - - r'. ‘ v t_.,r A - o ..~4 . . , ' ‘ - .' - .-1 ‘ . . a ‘ ' ' I . . ,v, ,.. ' " ‘ t:f_ . . _ . . 1 - , . - . ‘ ‘t I a - ml'.m1‘f,'u3q_l ' ... ‘ ui‘o . n - I ( .l -- IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII ' 33804090 LIBRAHY Michigan State University HMS & SUNS' 800K ”BINDERY INC. ABSTRACT LINKAGES BETWEEN T-GROUP TRAINER PERSONALITY VARIABLES AND TRAINER EFFECTIVENESS BY Sandra Kay Pinches There is little published data available concerning the role of the T-group leader in facilitating or impeding par- ticipant gains. Convergent research findings suggesting that all interpersonal behavior may be described by two dimensions, degree of self-acceptance/rejection and other-acceptance/ rejection (Foa, 1961), led to the hypothesis that construc- tive trainer behaviors communicate to participants high levels of both types of acceptance. Data was collected at an eight-day residential human relations laboratory sponsored by the State of Michigan Train- ing Laboratories in August, 1968. The 50 participants formed 5 T-groups, each led by a 2-trainer team. Participant self— acceptance (SA) and other-acceptance (0A) measures, distributed five weeks before and five months after the lab, registered participant gains, by T-group units. Near the end of the lab, members of each group rated their trainers on three effective- ness items, constituting a Quantitative Effectiveness (QE) :“I‘V‘A vvva .. ‘ | ”7" up Inn... IA‘N“ .\ All“; .I r \ ‘ x s. Sandra Kay Pinches score. Participants were also asked to comment in writing on which trainer behaviors were helpful and which were nohhelp— ful under each item, and to cite which trainer qualities were admired and which disliked. These comments were condensed into 50 perceived trainer personality traits (Pinches Vari- ables), and re-coded by a second judge, who agreed signifi- cantly (EI:1p experiences. theSe reports are performing some constructive functions in that they demand clarification of the group leader's role, qualifications and ethical reSponsibilities. Among the pro- fessional writers critical of confusion in this area is Gott- schalk (1966) who claims to have observed pathological reac- tions in members of National Training Laboratories groups at Bethel, Maine. Gottschalk attributed these emotional effects in part to lack of diagnostic and psychotherapeutic training in the backgrounds of group leaders. Whitman (1964), a group advocate, also interprets the leader's role within a psychotherapeutic framework. In Whitman's psychoanalytic theory, a successful group responds to the leader as a transference figure, and passes through a regressive experience during which neurotic behavior is exacer— bated, then resolved. Whitman believes that this process is best facilitated by a trainer who adOpts a non-directive, non- authoritarian approach. Benne (1964) also theorizes that the trainer's role is crucial in effecting constructive behavior change, but suggests that the most important group process involves role-modelling of trainer behavior. In Benne's View, a group becomes less defensive and more self-disclosing when exposed to a leader Who projects these qualities. A dissident view of the trainer's role is advanced by Gibb (1964), who argues that while the trainer's presence has an important impact on the group process, it is inhibitory rather than facilitative. Trainerless task-oriented groups in Gibb's study reportedly achieved maximum participative behavior and learning productivity more quickly than did the groups with leaders. One major deterent in evaluating these theories is the tendency of most writers to use vague descriptive labels which convey little Specific information about how trainers are behaving and how participants are responding. Compre- hensive, systematic descriptions of interpersonal behavior are necessary if linkages between trainer approaches and group responses are to be explored and effectively communicated. Convergent research findings from a number of studies suggest that all interpersonal behavior can be located and described on a circumplex structure with vertical and horizon- tal axes representing degree of self-acceptance/rejection and other-acceptance/rejection, respectively. Any particular interpersonal act may be plotted on the circumplex according to the amount of self-acceptance and amount of other-acceptance judged to be manifested in the act. The judges may include the self, significant others, and objective raters, depending on what level of information is sought regarding the actor's interpersonal behavior (Freedman, Leary, Ossorio, Coffey; 1951). The two axes section the circumplex into four quadrants, describing self-accepting/other-accepting, self- accepting/other-rejecting, self-rejecting/other accepting, and self-rejecting/other-rejecting behaviors. If a large number of observations are made on a single person, it is prob- able that all four quadrant descriptions will be used to describe his interpersonal actions. However, for any given individual, one of the quadrants is typically used more fre- quently than the other three, so that his personality may be described by a single quadrant label, depending upon the type of behavior he manifests most often (Adams, 1964). Adams (1964) points out that these four sector descrip- tions correspond roughly to traditional as well as modern clinical personality classifications. Among the earliest attempts at such typings are the four temperaments outlined by Hippocrates and subsequent philosophers. Thus a Sanguine temperament is characterized by a predominance of behaviors described by the self-accepting/other-accepting quadrant, while the remaining three regions correspond, respectively, to the Choleric, Melancholic, and Phlegmatic types. Berne (1966) theorized that the four possible combinations of self-acceptance/rejection and other-acceptance/rejection represent all the basic clinical classifications of healthy and psychopathological behavior. The position "I'm OK, you're OK" is the only possible constructive position, according to Berne. "I'm OK, you're not OK" corresponds to the clinical Paranoid classification, "I'm not OK, you're OK" to the depres- Sive position, and "I'm not OK, you're not OK" to the schizoid. Choosing the "I'm OK, you're OK" attitude over the other three, therefore, constitutes improved interpersonal behavior, or 9rOWth. It follows that if constructive behavior is defined by high acceptance of both self and others, helper behavior in this category will facilitate growth in the helpee, while helper behavior characterizing any of the other three inter- personal positions will have a destructive impact on the helpee. Bierman (1969) reviewed empirical data on the effects of the activity-passivity (self-acceptance/rejection) and other-acceptance/rejection dimensions on client change and child development. He cites research by Rogers (1967), Carkhuff and Truax (1966) and many others, showing that therapist positive regard and accurate empathic Understanding are directly related to independent measures of client out- come. Bierman quotes client-centered literature and research which suggests that the empathy dimension reflects relative levels of both other-acceptance and activity. Literature on therapist activity levels is reviewed according to various types of aCtivity. Research on therapist expressiveness is fairly consistent in showing that as therapists become more expressive and verbal, their clients follow suit. Bierman also concludes from his review that structured-directional activity on the part of the therapist, manifested by informa- tional statements and evaluative-prescriptions, exerts a posi- tive influence on therapy process and outcome. However, hostile-active behavior results in a restrictive effect on client expressiveness, self-exploration and motivation for therapy. Overall, Bierman's literature review supports the theory that constructive therapist behavior is characterized '9'! v- U‘ .8 I Rhi- v vale. _A 1 .. .Ow .. A it-..“ . “1 VA! vs“... -‘ru. [)1 [/7 p: ‘71 _ by relatively high levels of activity (self-acceptance) and other-acceptance, while therapist behavior characterized by low levels of either type of acceptance adversely affects the client's growth process. Foa (1961) describes several markedly similar grids and circumplexes constructed by modern researchers attempting to represent graphically the two interpersonal acceptance dimen- sions. One of the most elaborate of these is the Interpersonal Checklist (ICL) (LaForge & Suczek, 1955; Leary, 1956; LaForge, 1963), which was chosen as a T—group trainer assessment instrument in the present study. As shown in Figure l, the ICL circumplex classifies interpersonal behaviors into 16 categories (A-P), with a Dominance-Submission axis running vertically through categories A and I respectively, and a Love-Hate axis running horizontally through categories M and E. Included with Figure l are lists of personality traits corresponding to each of the categories A-P. The traits at the top of the list are the most heavily weighted and are located toward the circumference of the circumplex, while the traits at the bottom of the lists are least heavily weighted and located at the center of the circumplex. Adams (1964) argues that the Dominance-Submission and Love-Hate axes correspond to the self—acceptance/rejection and other—acceptance/rejection dimensions. In support of the same argument, Foa (1961) suggests that submissive behavior implies denial of affect or status to the self, although it li.‘nd.-u..§.|r 5 Eu) / .. , > x v......+.u.,-.\:\ / 63%. - III I: i All \4\\\ ,iflltu rflmJ SGlf effacing —MaSOChistic Figure l. Interpersonal checklist circumplex, and interpersonal checklist items by categories (cont'd on next page). Luv; “ 0516 H e R. e Q~ S 5. Rd E a 3'. :N c ‘1' 34 Av Av ab .1 v a a u. . E h» B Cu why P uh T 1 AD gnu 'Figure l (con't) Interpersonal Checklist Items by Categories N Tries to com— fort everyone Oversympathe- tic ~ Forgives any- thing Too lenient with others Encouraging to others Tender and soft-hearted Kind & reas- suring Considerate B Egotistical & conceited Boastful Somewhat snob- bish Proud & self- satisfied Independent Self-confi- dent Self-reliant & assertive Self-respect- ing 0 Spoils people with kindness Overprotec- tive to others Generous to a fault Too willing to give Bighearted & unselfish Enjoys taking care of others Good leader Helpful C Cold & unfeel- ing Thinks only of self Shrewd & cal— calculating Selfish Businesslike Can be indif- ferent Likes to com- pete with others Able to take care of self 3 Expects every- one to admire him Tries to be too successful Acts impor- tant Always giving advice Often admired Makes a good impression Respected by others Well thought of 2 Cruel & unkind Sarcastic Self-seeking Impatient with others' mis- takes Stern but fair Hard-boiled when needed Firm but just Can be strict if necessary A Dictatorial Bossy Dominating Manages others Forceful Good leader Likes responsi- bility Able to give orders E Hard-hearted Frequently angry Outspoken Often unfriendly Irritable Critical of others Straightforward and direct Can be frank & honest E Rebels against everything Bitter Complaining Resentful Skeptical Often gloomy Resents being bossed Can complain if necessary .3. Clinging vine Hardly ever talks back Dependent Wants to be led - Admires & imi- tates others Often helped by others Very respect- ful to author- ities Grateful Figure l E Distrusts everybody Jealous Stubborn Slow to for- give a wrong Hard to im— press Frequently disappointed Touchy & easily hurt Able to doubt others .15 Will believe anyone Easily fooled Likes to be taken care of Lets others make decisions Accepts advice readily Very anxious to be approved of Trusting & eager to please Appreciative (con't) E Always ashamed of self Timid Self-punishing Shy Easily embar- rassed Lacks self— confidence Apologetic Able to criti- cize self .1; Agrees with everyone Wants every- ones' love Will confide in anyone Too easily in- fluenced by friends Wants everyone to like him Always agree- able & pleasant Eager to get along with others Cooperative I Spineless Meek Passive & unaggressive Obeys too willingly Easily led Modest Easily gives in Can be obedi- ent M Loves every- one Fond of every- one Likes every- body Friendly all the time Warm Sociable & neighborly Affectionate & understanding Friendly a YA : ~ nub . 1 a .P m .AF» ’1‘ . ‘4 Qt . CH o. N‘ h'f' 5" b M 56; 1E2; a.. 10 may admit giving acceptance to others. Dominance, on the other hand, implies social acceptance of self, giving the self status, even if the other person is socially rejected in the process. Foa notes that behavior descriptions such as "self-satisfied", "self-respecting" and "self-confident" occupy the Dominance hemisphere of the circumplex, while "self-punishing" and "ashamed of self" occupy the Submissive hemisphere. The Love-Hate dimension more obviously describes a positive-negative polarity in the actor's behavior towards others. Personal growth or positive change may be defined as an increase in behavior described as Dominant (self—accept- ant), and an increase in behavior described as Loving (other— acceptant). In terms of the ICL quadrants, improved inter- personal behavior is characterized by a greater number of acts classified as Dominant—Loving and a lesser number of acts classified in the other quadrants. Research on interpersonal impact using the ICL quadrant method included Mueller (1969) who coded behaviors of psycho- therapists and their clients into the ICL circumplex, using audiotapes of individual therapy sessions. Mueller labeled the quadrants as competitive-hostile (categories BCDE), passive-resistant (FGHI), support-seeking (JKLM) and supportive- interpretive (NOPA), to make them more obviously applicable to therapist and client roles. He then evaluated the recipro- cal impact of clients and therapists by scoring the frequency of various possible response sequences. For example, the ”u IE *¢ 6. 5 ~\w ‘1.” \\H 11 therapist might make a statement scored NOPA and the client's response might be scored FGHI. Mueller concluded that certain types of behavior show a powerful tendency to elicit corre- sponding responses in the other interactant, and these results clarify the nature of transference and countertransference in terms of reciprocal interpersonal impact of therapist and client. In a related study using the same data pool (the Michigan State University Counseling Center Tape Library), Crowder (1972) correlated therapist behavior, described by Mueller's (1969) quadrant labels, with constructive client change measured by the MMPI. Crowder's hypothesis was that suppor- tive-interpretive behaviors (NOPA) are appropriate to the psychotherapist role, while therapist responses scored in other quadrants constitute countertransference. However, Crowder's results only partially supported this hypothesis. Therapists who had the most constructive impact were dis- tinguished from unsuccessful therapists by their positions on only one dimension, the Dominance-Submission axis. Although the majority of acts of all the clinicians in the study were scored as supportive-interpretive, successful therapists were more often dominant in their interview behavior, while less successful therapists were more submissive. In summary, although little research has been published clarifying which T-group trainer behaviors are linked with Participant gains, hypotheses may be drawn from other research H- o - ."" o , ‘aor OI'V‘ : “A ‘ thy :Fnav . a “Huh-n a N "".h .‘ (I) 12 in interpersonal behavior, particularly psychotherapy. Theory and empirical data in the area of helper-helpee inter- actions consistently suggest that the two dimensions self- acceptance/rejection and other-acceptance/rejection account for helper constructiveness and helpee gains. Specifically, positive changes in the helpee's interpersonal behavior are typified by increased acceptance of self and others. Helpful behavior consists of a predominance of acts manifesting high levels of self and other acceptance, while behavior manifest- ing low levels of either type of acceptance may be destruc- tive to the helpee. The purpose of the present study is to investigate the relationship between T-group trainer behavior as perceived by group members, group-rated trainer effectiveness, and inde- pendent measures of participant gains on self-acceptance and other-acceptance. Hypotheses Participants' feedback on helpful and blocking trainer behaviors are analyzed to develop trainer personality cate- gories, which are then used to score each trainer on the ICL Dominance-Submission (Dom) dimension and the ICL Love-Hate (Lov) dimension. These scores determine the trainer's plotted positions on the ICL circumplex, which facilitate comparison with correSponding plottings of mean net T-group changes on self-acceptance (SA) and other-acceptance (OA). An independent . b «q, 31.. p: (I) r‘l l3 trainer effectiveness score (QE), derived from ratings of trainers by their group members, is correlated with trainer ICL scores and also shown on the graphical presentation described above. It is hypothesized that trainers rated as most effective are most Dominant and Loving, while trainers rated as least effective are least Dominant and Loving. It is also hypothesized that T-groups who perceive their trainers as Dominant and Loving show mean net gains on SA and CA. Since in the present study each T-group was led by a team of two trainers, it is hypothesized that mean group changes are more closely related to combined trainer scores on Dom, Lov and QE than to individual trainer scores. The trainer team effectiveness score was called Overall Trainer Effectiveness (OTE). Finally, the participant feedback on trainer behavior is examined for more specific linkages with QE. In summary, the following hypotheses are generated: (1) Graphically, trainers with high QE scores are described by their groups as occupying the Dominant-Loving quadrant of the ICL circumplex. Mean group net changes on SA and CA reflect the trainers'Dom and Lov scores; the higher the trainers' scores on these dimensions, the greater the positive changes on SA and CA. (2) Dom correlates positively with QE. (3) Lov correlates positively with QE. 14 (4) XDom for trainer teams correlates positive with OTE, SA and DA. (5) ZLov for trainer teams correlates positively with OTE, SA and 0A. (6) Linkages exist among more specific trainer behaviors and QE. METHOD Gathering of Data The data analyzed in the present study was collected in August, 1968, as part of a comprehensive project investigating the nature of participant gains and the role of trainers in a human relations laboratory (Force, 1969; Hurley & Force, 1972). The eight-day residential laboratory at which the design was implemented was sponsored by the State of Michigan Training Laboratories, a university affiliated organization which then conducted such training programs semi-annually. The explicit goals of the August, 1968, lab involved gaining a clearer grasp of one's strengths and limitations in interpersonal communication skills, with special attention directed to con— structive encounters, feedback process, and the transfer of new learning to "back home" settings. The 33 male and 17 female participants included 13 junior and senior high school teachers, 11 graduate students in social work or psychology, 5 school principals, 5 pastors or priests, 4 school counselors, 2 each of housewives, professors, social case workers and school superintendents, 1 curriculum consult- ant, l psychiatrist, l director of marketing and 1 art co- ordinator. Five T-groups of equal size were formed, with two 15 16 trainers assigned to each group. In the initial phase of the laboratory, participation in these groups received primary emphasis; later, the 10 member groups were divided into pairs of helping partners, and finally, pairs selected from separate groups coalesced into trainerless sextets. The daily time schedule devoted about nine hours to laboratory activities. About 50% of the total time was spent in T-group participation, 30% in sextet sessions, 15% in cognitive session, including lecturettes, and about 5% in research participation. The trainers varied in experience and background and were designated as "senior" trainers and "junior" trainers, mostly on the basis of these variables. The five senior trainers were all PhD's, including one counseling psychologist, one social psychologist and three clinical psychologists. Four senior trainers and one junior trainer had completed eight-week summer internships in sensitivity training at Bethel, Maine, under the auspices of National Training Laboratories. The five junior trainers included two PhD's, one in clinical psychology and one in educational administration. The remain- ing three, educated to at least the MA level, were a clergyman, a high school counselor and a graduate student in social psychology. One senior trainer and one junior trainer co-led each T-group. To assess participant growth, data packets containing 10 personality variable measures were mailed to each participant 17 five weeks before and again five months after the laboratory. The same packet was mailed to one intimate and one job colleague of the participant's choice. Seven variables con- stituted a "self-acceptance" (SA) cluster, measuring Openness, data-seeking, data-giving, how "OK" the participant regards himself (Berne, 1966, p. 270), dominance-submission (LaForge & Suczek, 1955; LaForge, 1963), power and effectiveness in work, and activity and expressiveness (Harrison & Oshry, 1965). The three remaining variables, forming an "other-acceptance" (OA) cluster, included how "OK" the participant regards others (Berne, 1966), interpersonal warmth and acceptance (Harrison & Oshry, 1965), and love—hate (LaForge & Suczek, 1955; LaForge, 1963). Pre-lab and post-lab scores for each participant were calculated by summing his self-rated score on each variable with the rating submitted by his colleague and intimate, then summing these totals across the seven SA variables. Individual OA scores were calculated in the same manner, summing across the three OA variables. Net change scores on SA and 0A for each participant consisted of the difference between his pre- lab scores and his post-lab scores on the two acceptance dimensions. Finally, a mean net change score was calculated for each of the five T—groups. As shown in Table l, Hurley and Force (1972) found that groups 2, 3, and 5 registered mean net gains Ohithe SA measure, while groups 1 and 4 registered losses. On the OA measure, groups 2 and 5 registered mean net gains, while groups 1, 3 and 4 registered losses. 18 Table 1 Mean Net Changes on SA and OA by T-group Units T—Group No. Change on SA Change on OA 1 -O.22 -l.92 2 7.39 2.90 3 3.81 -1.10 4 -1.65 -O.86 5 8 18 2.38 To help assess how the trainers facilitated or blocked participant gains in interpersonal competence, group members were asked near the end of the laboratory to independently rate the effectiveness of each of their T-group trainers. In two similar laboratories held in March and August, 1967, participants had rated trainers as very effective (3), quite effective (2), somewhat effective (1), or not effective (0), on the following four items: A: This trainer helped me to recognize conditions relating to the operating effectiveness of our T-group. B: This trainer did things in the group that enhanced learning and change. C: This trainer introduced and implemented alternative ways of problem-solving. D: This trainer presented ideas and concepts in a way that was useful. 19 In the August, 1968, lab, the same rating scale was applied to the following revised three items: A: The trainer acted in ways which helped the T— roup to be more effective. B: The trainer helped me to become more aware of my personal hang-ups and of ways in which I might change my behavior. C: The trainer demonstrated understanding of me as an indi— vidual. A Quantitative Effectiveness (QE) score was calculated for each trainer by summing the total number of points assigned on each item, dividing the result by the number of group members contributing ratings, then summing the means for the three items. Since the participant gain scores represented means for five T-groups, mean trainer team scores on the effective- ness instrument were calculated from the individual trainer QE scores and labeled Overall Trainer Effectiveness (OTE). The QE and OTE scores for the August, 1968, lab are shown in Table 2. Table 2 Effectiveness Scores for Individual Trainers and Trainer Teams Trainers QB Trainers QB Teams OTE 1 59 6 69 1&6 6.39 2 75 7 66 2&7 7.13 3 71 8 68 3&8 6.95 4 70 9 53 4&9 6.16 5 75 10 70 5&10 7.25 20 Hurley and Force (1972) established that the QE score was a robust and reliable measure of trainer effectiveness. QE scores and rankings of trainers on this variable were available for four trainers from the March and August, 1967, laboratories. One such trainer ranked first out of 10 trainers in August, 1968, second of 11 and second out of 12 in the previous two laboratories. A second person ranked fifth out of 10, fourth out of 11 and fourth out of 12, while a third ranked sixth out of 10, third out of 12, and first out of 11. The last trainer for whom data was available ranked third out of 10 in 1968 and sixth out of 12 previously. The absolute value of scores was also stable over time in Spite of changes in item content and group membership. Within a functional range of from 5.7 to 11.7 points, two particular trainers were rated less than .1 point apart in two laboratories held six months apart, while another trainer's score varied by only .4 point across laboratories. Finally, Hurley and Force (1972) conclude that up to 96% of participant gains by T-group units, measured five months after the laboratory with the previously described data packets, were related to how effective members had rated their trainers on the QE scale. Beyond supporting the use of the QE score as a measure of trainer effectiveness, these results emphasize the critical role played by effectiveness in deter- mining the direction and magnitude of change in participants. "h (J l l K tr ..E ’(J ['1’ (‘l 21 In addition to assigning numerical ratings on the three QE items, group members were also requested near the end of the lab to write some descriptive statements (Qualitative Data) of how the trainer had been helpful or blocking in each of these areas (see Appendix A). During this phase a fourth item (D) was also included, asking participants to list which trainer personality qualities they would like to acquire and which they would prefer to avoid. These Qualitative Data served the present study as a source of information on how participants perceived their trainers' behavior. Perceived trainer behaviors could then be linked with QE rankings, and trainers' personalities could be plotted on the interpersonal circumplex described previously. In summary, previous analyses of data collected at the August, 1968, human relations laboratory together with find- ings from earlier laboratories have led to several conclu- sions. First, participants showed changes in self-acceptance/ rejection (SA) and in other-acceptance/rejection (OA) over a six-month interval beginning five weeks before the lab. Second, trainer effectiveness (QE) as rated by the participants themselves was firmly linked to both positive and negative participant changes of the T—group units. Third, trainers' rankings on this effectiveness dimension proved to be stable over time in spite of revisions in the items rated, changes in staff composition and changes in T-group membership. nc whj wi1 51C tr ‘31 t: E) (I) “J 1r 22 Analysis of Qualitative Data Participants' written descriptions of "helpful" and "nonhelpful" trainer behaviors were analyzed to determine which perceived trainer personality variables were associated with independent measures of trainer effectiveness and par- ticipant gain. The feedback form (Appendix A) provided separate columns for written comments on facilitative or blocking behavior relevant to each of the four effectiveness items (A-D). Initial inspection of the Qualitative Data revealed that several of the personality qualities of trainers cited by participants recurred frequently across group members and across trainers. It was thought that a list of personality traits might be developed which would include the most fre- quently mentioned trainer behaviors. Linkages between trainer effectiveness and personality variables could then be explored more effectively, by condensing the written feedback as much as possible into scores on such variables. This pro- cedure resulted in a list of trainer personality traits which seemed heavily weighted in the direction of dominant trainer behaviors. Since in the ICL system, Dom is calculated by sub- tracting scores on submissive categories from dominant cate- gories, it was decided that the initial list of trainer variables would result in inflated Dom scores, should the ICL description be applied to them. Therefore, the list of cate- gories (Pinches Variables) was expanded to include virtually 23 every trainer personality trait mentioned more than once in the written feedback. Interpretation and accompanying dis- tortions were reduced to a minimum by adding another variable description for any comment which did not seem to fit easily into the existing categories. In the process of scanning the written comments, it was noted that certain behavioral descriptions of particular trainers were mentioned in the "helpful" column by some of their group members and again in the "blocking" column by other participants. In some cases, a group member listed identical trainer behaviors in both columns. The question arose as to whether participants were using the same labels to describe different behaviors, or whether various group members might have opposite emotional reactions to the same trainer behavior. It was noted further that some partici- pants had such strong positive feelings about their trainers that they wrote only positive comments, while others had equally strong negative feelings and wrote only criticisms of their trainers. Participants' liking for their trainers may have led them to classify most or all of perceived trainer behavior as "helpful", while dislike may have led group members to attack all trainer behavior as destructive. Although the participants' ratings of trainer effectiveness (QE) on the three broad areas of trainer behavior (A-C) were shown by Hurley and Force (1972) to correlate impressively with more objective measure of participant gains (SA and 0A), trusting CCI‘LE ati 24 participants judgments about the sources of effectiveness was considered to involve too many blind assumptions. For example, a participant variable such as characteristic response to negative feedback could lead one group member to view trainer negativity as a helpful confrontation, and another group member to view the same behavior as a non-helpful attack. Nevertheless, both participants might conceivably experience some feelings as a result of negative feedback, leading to self-exploration, increased interaction with others, and growth. Because the influence of participants' liking for a trainer on the written feedback was unknown, and because the present study was exploratory in nature, it was decided to ignore the "helpful" versus "nonhelpful" column membership in assigning written comments to categories. Exceptions were made in cases where column membership obviously affected the meaning of a comment; this occurred most often when a blank space was left under one or the other column. A blank or equivalent remark indicated either that the participant had no positive observations or no negative observations to write, and thus column membership differentiated two opposite evalu- ations of the trainer. The written comments submitted by the participants were ultimately condensed into 50 perceived trainer qualities (Pinches Variables), including 47 personality traits, a cate- gory for comments judged not to fit the previous 47 descrip- tions, and two categories scoring the number of members who 25 omitted written comments in the "helpful" and "nonhelpful" columns (see Appendix C). A trainer's score on any Pinches Variable was calculated by summing the number of group members who attributed this trait to him. This method helped insure that the final descriptions of the trainers' person- alities were those of "generalized group members" (Freedman, Leary, Ossorio & Coffey; 1951), and idiosyncratic perceptions were kept in perspective. The total number of comments scored ranged between 47 and 82 for the 10 trainers, includ- ing between 2 and 13 comments per trainer which did not fit the Pinches Variable descriptions (see Appendix D). Reliability of Pinches Variables To check the reliability of the coding into Pinches Variables, the comments in the Qualitative Data scored as units by Pinches were marked on the feedback sheets, and every fourth unit was selected for co-rating by a second judge. From the resultant total of 251 comments, two sets of 60 units each were chosen, using a random number table. These sets were scored by the co-rater in two training sessions, and the results were not included in the final reliability statistic. The second judge was an undergraduate senior who had participated in various T-groups and had also taken course- work on the theory and practice of sensitivity training, encounter groups and related group work. Before the first trial scoring she was given the instructions shown in 26 Appendix B, a set of scoring sheets, the Qualitative Data with the selected units underlined, and additional verbal instruc- tions emphasizing the goal to simply condense rather than interpret the data. After completion of the first set of co- ratings, the two judges conferred about comments they had assigned to different categories. A principle source of dis- agreement was a tendency on the part of the co-rater to interpret group members' observations on the basis of her own feelings about what constituted helpful trainer behavior. The instructions to condense rather than interpret the group members' comments were strongly reiterated, and the second set of training comments were delivered to the co-rater. Following completion of this set, disagreements were again discussed and minor clarifications of the instructions were made. The final set of 131 co-ratings included 92 agreements with the original coding, 36 disagreements and three comments omitted by clerical errors of Pinches or the co-rater. In order to calculate a Chi-Square, it was first necessary to estimate the expected frequency of agreement of two judges, given 50 categories with unequal chances of being selected. An appropriate statistic, Fe, is described by Scott and Wertheimer (1962): Ee=Bl First a frequency chart is constructed as shown in Table 3, with each judgment placed in a cell according to the coding 27 Table 3 Frequency Distribution of Comments Assigned To Pinches Variables by Co-raters, Column Sums, Sums Squared and 3e 7 Judge 2 2A 2B 2C 3 4 EA 5B 5C 5D 6A 6B 6C Judge 1 ran—- 2A 2 28 2 1 2C 5A 3 1 5B 1 5C 2 5D 6A 1 1 8 6B 1 6C 6D 6E 6F 6G 1 l 10 11 1 12-28 29 l l 30 31 32 33 34-38 1 Sum 2 2 5 1 1 8 3 1 2 l 8 0 0 Sq. 4 4 25 1 l 64 9 l 4 1 64 O 0 Judge 1 2A 2B 2C 4-5D 6A 6B 6C 6D 6E 6F 6G 10 11 12 13 14 15 16-25 26 27-38B Sum Sq. 28 Judge 2 .51 2 l l O l 0 Table 3 (cont'd) 10 11 12.12.12 1 1 1 1 O 2 1 0 4 Judge 1 l-SB 5C 5D 6A-6G 8-16 17 18 19 20 21 22 23 24 25 26 27 28 29-36 37 38A 38B Sum Sq. 29 Table 3 (cont'd) Judge 2 2.1.22. 1 l l 2 1 4 l 23 24 64 2.5.2.6. 1 3 1 4 1 l6 1 l6 16 Judge 1 1 2A 23 2C 3 4 5A SB 5C 6A 6B 6C 6D 6E 6F 6G 7-23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38A Sum Sq. P _e 5 25 2 4 1 l 30 Table 3 (cont' Judge 2 2.2.2.225. l 3 3 l 0 9 l 0 5 agreements/128 cases d) 36 37 38A 38B 1 17 1 17 1 289 31 labels assigned to it by each judge. The frequencies given by the column sums are divided by N, in this case 128, to give the mean Bi for each column or category. These numbers give the proportion of cases which would be coded into the various categories by chance. To obtain the probability of agreement of two judges by chance, these numbers are squared. Finally, 2: is summed across the columns to yield 23:, or chance agreement for two judges. As shown in Table 3, Be for the Pinches Variables is about five agreements out of 128 judgments. The Chi—square reliability statistic for the co-rating of the Pinches Variables is shown in Table 4. Table 4 Chi-square Testing Reliability of Co—rating of Pinches Variables Expected Observed FrequenCies FrequenCies Judges + - §EE§ Judges + - Sums l 128 0 128 l 128 O 128 2 5 123 128 2 92 36 128 Sums 133 123 256 Sums 220 36 256 2 x = 3027.6 (p_< .001. 91: = 1) In the contingency tables, the original judgments made by Pinches are labeled positive when they agree with Pinches and 32 negative when they disagree. For an expected frequency of agreement of five judgments and an observed frequency of agree- ment of 92 judgments, Chi—square equals 3027.6, which consti- tutes highly significant agreement (p<1.001, d: = 1). LaForge Circumplex A total of four judges independently assigned ICL cate- gory descriptions to the list of Pinches Variables. The judges included Pinches, John R. Hurley, William J. Mueller and Barbara Brandt. Hurley was one of the ten trainers who par- ticipated in the August 1968 lab, helped design the research program implemented there, and was very familiar with the history and application of the ICL in interpersonal research. Mueller has supervised or conducted much of the research at Michigan State University on psychotherapy process and thera- pist variables as described by the ICL. Barbara Brandt was one of the graduate students trained by him to code audio- taped segments of psychotherapy sessions into the 16 ICL categories. Hurley and Pinches considered themselves to be more familiar with the Qualitative Data and the meanings of acts occurring within the T-group context than were the other two judges, so discrepancies in judgments were resolved by a conference between them. A final list of category assignments was developed during this conference (see Appendix D). Scores on the Pinches Variables were summed across all the variables assigned to each ICL category, for each trainer. From these, 33 the two summary dimension scores, Dominance-Submission (Dom) and Love-Hate (Lov) were calculated for the trainers. Nine Pinches Variables were considered by one or more judges not to fit the ICL descriptions and were therefore excluded from assignment to circumplex categories. These included the following variables: 6B, 28, 29, 30, 31, 36, 37, 38A, and 38B (see Appendix C). Reliability of ICL assignments. The reliability statis- tic employed was Dittman's R, which was used by Mueller (1969) in applying the 16 ICL category descriptions to therapist and client responses in individual counseling sessions. 3 is calculated by 16 categories using the formula: n B 1-2 EMU/4 i=1 Dittman's g is the product of the amount of disagreement be- tween judges, ranging from zero to eight categories, times the number of units for which disagreement occurs. This product is summed over the nine possible levels of discrep- ancy and divided by the total number of units, in this case 41, to yield the numerator of the above formula. R’is then converted to E by applying the formula: 35 = 1.706 R As shown in Table 5, all paired comparisons among the four judges were highly significantly (E‘7°001I.§£ = 40) on the category assignments. 34 Table 5 Interjudge Agreement on Assignment of 41 Pinches Variables on ICL Categories Judges # Cate— 1&2 1&3 1&4 2&3 2&4 3&4 gories # # # # # # Disagree Units g Units g Units 6 Units g Units 0 Units g 0 20 0 2O 0 l3 0 14 0 ll 0 16 0 1 10 10 9 9 12 12 13 l3 l6 16 10 10 2 5 10 2 4 6 12 7 14 4 8 7 l4 3 3 9 5 15 0 0 l 3 2 6 2 6 4 2 8 2 8 3 12 5 20 5 20 2 8 5 l 5 3 15 4 20 l 5 1 5 2 10 6 0 0 0 0 2 12 0 0 l 6 1 6 7 0 0 0 0 l 8 0 0 l 8 l 8 n 2 g 42 51 76 55 69 62 i=1 R .745 .665 .690 .580 .538 .622 E 8.14* 7.26* 7.54* 6.34* 5.88* 6.79* *B< .001, g; = 40 Correlations. Product-moment correlations were determined among the entire body of variables describing the 10 trainers, including the 50 Pinches Variables, the 16 ICL category scores, the 2 ICL summary scores, and the QE ratings. The participant gain scores represented mean scores for the trainer teams also. The mean QE score (OTE) had previously been determined by Hurley and Force (1972). Trainer team 35 summary ICL scores were derived by adding each trainer's score to that of his co-trainer on each acceptance dimension. Product-moment correlations were determined among the trainer ICL scores, OTE, and the mean participant gain scores on SA and CA. RESULTS Hypothesis 1 Table 6 shows the ICL Dom scores for the 10 individual trainers, ranging from 21.1 to 41.0, and their ICL Lov scores, ranging from -2.3 to 19.0. No trainer registered a negative score on the Dom dimension, while only one trainer registered a negative Lov score. Table 6 ICL Dom and Lov Scores Trainers 29m Lg! Trainers 22m Egg 1 21.1 4.4 6 25.9 6.4 2 31.7 5.6 7 21.5 19.4 3 30.4 15.6 8 31.4 7.1 4 26.4 2.0 9 23 2 9.7 5 41.0 -1.9 10 33.0 4.8 As shown in Figure 2, all 10 trainers were described by their group members as occupying ICL sector NJA. The three trainers with the highest QE scores (see Table 2) were clus- tered in the two categories A and P, and the three lowest scorers on the QE variable were distributed among categories A, P, and O. 36 37 DCM4 1 4O oTlO T213 0T8 27T3 0T4 0T6 AT9 ,ATl 11T7 SLHB Key [HT Most effective trainers .AT Least effective trainers 0G T-Groups Figure 2. Positions of ten trainers on ICL circumplex, showing relative trainer effectiveness, and heads of net T—group change vectors. 38 The three most effective (QE) trainers were also among the four highest scorers on the Dom dimension, while the three trainers rated as least effective received the three lowest Dom scores. Only one trainer (10) deviates from the pattern, ranking second on Dom, and fourth on QE. This devi- ation is, however, too slight to alter the pattern of results. The positive relationship between relative Dom and relative QE strongly supports Hypothesis 1. Predictions regarding the influence of Lov on QE were not fulfilled. The Lov scores of the three most effective trainers ranged from -2.3 for Trainer 5, to 15.2 for Trainer 3. The least effective trainers occupy an overlapping range in Figure 2, with Trainer 1 at 4.0 on the Lov dimension, and Trainer 7 at 19.0. Therefore, Hypothesis 1 was supported only in part with respect to the relationship between the two acceptance dimensions and trainer effectiveness. Although QE varies directly with trainer Dom, trainer Lov is not obviously linked to QB in the graphic presentation. The positions of the T-groups in Figure 2 represent the heads of mean net change vectors plotted from the origin, using the SA and 0A scores (see Table 1) calculated by Hurley and Force (1972). Groups 1 through 5 were led by senior trainers 1 through 5 and junior trainers 6 through 10, respectively. Groups 2, 3 and 5, whose trainers were among the four highest scorers on Dom, increased on SA. Their corre- sponding junior trainers also scored high on Dom, except for 39 Trainer 7, who registered one of the lowest Dom scores. Groups 1 and 4, whose senior and junior trainers clustered at the lower end of the Dom rankings, decreased on SA. No clear pattern emerges in Figure 2 regarding the rela- tionship between trainer Dom and participant gains on CA. The pattern relating trainer Lov to mean net group change on SA and CA is also unclear from this Figure. Groups 2, 3, and 5, who gained on SA, were led by trainers whose positions on the Lov dimension span the entire range of trainer Lov scores. The trainers leading groups 1 and 4, who decreased on SA, registered a smaller but overlapping range on the Lov dimension. The same overlap in trainer Lov score ranges is shown for the trainers leading Groups 1, 3, and 4, who decreased on CA, and those leading Groups 2 and 5, who gained on OA. With respect to the overall hypothesis that trainers exert a "pull" toward their own ICL quadrant position, an interesting pattern is shown in Figure 2. Although all the trainers were described as Dominant-Loving (N-A), Groups 2, 3, and 5, led by the most Dominant trainers, moved more deeply into the same quadrant, while Groups 1 and 4 moved in the opposite direction. In terms of the 16 ICL categories, Group 1 showed a slightly stronger tendency to complain and rebel (E,F), and Group 4 became slightly more submissive and self- critical (H,I). Again the trainers leading Groups 1 and 4 were among the least Dominant trainers, but were not different 40 from other trainers on the Lov dimension. Furthermore, the difference in the responses of Groups 1 and 4 cannot be accounted for by differences in the quadrant positions of their respective trainers. Hypothesis 2 Because of the directionality of the following hypothe- ses concerning the linkages between trainer ICL scores, QE, SA and OA, and because of the small sample sizes involved, it was decided to use one-tailed tests. Hypothesis 2 was supported; Dom X§° QE = .73 (pf<.05, df = 8, l-tailed). This correlation is consistent with the conclusion drawn from Figure 2, that trainer Dom is clearly linked with trainer effectiveness. Hypothesis 3 Hypothesis 3 was not supported; Lov XE: QE = —.26 (pg). This finding is also consistent with the conclusion drawn from Figure 2 that trainers' positions on the Lov dimension were not related to relative trainer effectiveness. Hypothesis 4 Correlations among the combined trainer acceptance scores ZLov and ZDom, OTE, and participant gain scores SA and CA are shown in Table 7. The variables labeled A, B, and C refer to the component items of the OTE (QE) rating instrument. 41 Table 7 Correlations Among Trainer Team ICL Scores, Overall Trainer Effectiveness and Participant Gains ZDom ZLov Z(Dom+Lov) OTE .64 .28 .83* A .65 .ll .70 B .80 .15 .88** C -.24 .93** .51 SA .26 .28 .81* 0A .51 .04 .52 * 2<-1o. 2: ** p<.05, df 3, 1—tai1ed 3, l-tailed The acceptable significance level for correlations was in- creased from .05 to .10 because of the decrease in sample size from an N of 10 individual trainers to an N of 5 T-groups or trainer teams. Hypothesis 4 was not supported. ZDom did not correlate with SA, OA, OTE, nor with any component of OTE. Hypothesis 5 Hypothesis 5 was supported in only one respect. Combined trainer Lov (ZLov) did not correlate significantly with Sa, OA, or OTE, but ZLov did correlate significantly with OTE item C, ("Showed understanding of me as an individual"). ZDom correlates much more highly than does ZLov with all the 42 variables in Table 7, except for item C, where the relation- ship is reversed in favor of ZLov. Further inSpection of the combined trainer acceptance scores suggested that an overall acceptance score, summing the independent Dom and Lov scores, might be more closely related to OTE and participant gain scores than either dimen- sion taken alone. Overall acceptance scores (Dom+Lov) were developed for the 10 individual trainers by adding each trainer's Dom score to his Lov score. A combined overall acceptance score, Z(Dom+Lov), was then determined by summing the overall acceptance scores of the two trainers in each team. The overall acceptance score for individuals is pre- sumed to measure the total amount of acceptance perceived in each trainer by his group members. The combined overall score reflects the total amount of acceptance perceived in the trainer team, theoretically accounting for all perceived con- structive influences by the trainers. The overall scores for individual trainers were correlated with QE; (Dom+Lov) Xfi' QE = .47 (as). The results for the combined overall scores are shown in the third column of Table 7. As was hypothesized, Z(Dom+Lov) is significantly correlated with more variables than is either ZDom or ZLov. The combined overall score corre- 1ates with OTE, OTE item B ("Helped me become more aware of my hand-ups and ways I might change"), and SA. 43 Hypothesis 6 Correlations between individual trainer variables and QE were analyzed with the aim of pinpointing more specific sources of trainer effectiveness than the ICL acceptance dimensions. Because of the exploratory nature of this phase of analysis, no directional hypotheses were formulated, and two-tailed tests of significance were used. First the correlations between the 16 ICL categories and QE were examined; none even approached the .05 level of significance. Next, a typal analysis of the Pinches Variables was made using the McQuitty Elementary Factor Analysis method (McQuitty, 1961). One modification was made in the original list of variables before the correlations were determined; Variable 8 ("Evaded confrontation") was collapsed with Variable 9 ("Deferent to authority"). This procedure was adopted because Variable 9 was mentioned only once in the Qualitative Data, in the context of a junior trainer evading confrontation with his senior partner. Therefore, Variables 8 and 9 appear in the matrix as (8+9). The McQuitty method is to first circle the highest corre- lation in each column of the matrix, and to select from these the highest entry in the matrix. The two variables mediated by the highest entry constitute the core members of the first type. Pinches Variables l6 and 18 met this criterion. Next, reading across rows 16 and 18 of the matrix, all the circled column entries falling in these rows are added to Type I. 44 These are variables which are more closely related to 16 and 18 than to any other variables. The rows of the added vari- ables, in this case 6C and 21, are examined in the same manner for other circled entries. When no more members are found, Type I is exhausted and the next highest matrix entry is selected for the core of Type II. Exhaustive analysis of the Pinches Variables yielded 14 Types, the core variables of which were all significantly correlated (p< .05 or pf:.01, gf = 8, 2-tailed). Table 8 lists the descriptions of the variables included in the 14 Types and designates which vari— ables correlate negatively with other members of their clusters. Factor loadings are obtained by constructing submatrices composed of the variables of each Type. The columns are summed, and the variable with the highest total is selected as reference factor for that type, designated RFC. The refer- ence factors for each Type are designated in Table 8 by an asterisk. The factor loadings for RFC'S are the column entries of the corresponding Pinches Variable in the original matrix. Table 9 shows the 14 reference factors and their loadings on the Pinches Variables. Although the 14 Types are statistically good clusters internally, with significantly high inter-correlations among member variables, and also good externally, with relatively few significant correlations with other types or members of other types, the content analysis reveals less clear patterns. TYPe II III IV 45 Table 8 McQuitty Type Descriptions, Designating Negatively Correlated Variables (—) and Reference Factors (*) Var. No. Des. Variable Descriptions l6 Smug, self-satisfied 18 * Capable of self—criticism 21 Direct, outspoken, straightforward 6C Let members work out own solutions 6G Could be supportive 17 Defensive, felt attacked 36 * Inconsistent, double-binding, con- fusing 5B Abrupt, jumped the gun, out people off 19 Open, self—disclosing 26 (-) Cool, aloof 34 (-) Trustworthy, not-threatening 6D Suggestive, gave advice 35 * Sometimes refused to support 3 Questioned deeply and competently 2B High amount of contact with indi- viduals 8+9 Deferent to authority 24 * Involved, caring, sincere, interested 13 (-) Perceived as "group member" 5D Made value judgments 20 * Frank, honest, straight 7 Very confronting 46 Table 8 (cont'd) Type Var. No. Des. Variable Descriptions VI 1 (—) Paid attention to group process 32 Good listener 38B * Received relatively little negative feedback 25 Uninvolved, not caring 14 Non-directive at times 29 Understanding 6A Gave accurate, clarifying feedback VII 2A (-) Paid too little attention to indi- viduals QE * Quantitative Effectiveness Score 10 Perceived as authority figure VIII 6E Disproportionately negative 15 * Self-confident, non-defensive IX 6B * Played favorites 2C Paid too much attention to individuals X 30 (—)* Not understanding 4 Brought people out 38A Received relatively little positive feedback XI 5A Impatient, bored 23 * Spontaneous XII 28 * Non-verbally sensitive 33 Accepting, non-judgmental 27 Warm, friendly 12 Perceived as "trainer" Type XIII XIV Var. 5C 22 11 6F 31 NO. 47 Table 8 (cont'd) Des. Variable Descriptions * Attacking, hurtful (-) Playing a role (—)* Perceived as strong, helpful Could be negative Used alternative methods well 48 Table 9 McQuitty Reference Factors and Loadings on Pinches Variables** Vars. RF. RF.. RF... RF. RF RF . RF .. RF ... RF. -—1 -—11 ——111 -—iv —-V ——V1 ——v11 ——V111 ——1x 16 88* -25 32 -48 62 -26 38 -10 30 18 -- -35 70* -50 39 -23 41 -l4 19 21 80* -44 44 -32 42 -64* 22 -33 -15 6C 72* -16 63 14 13 06 27 -19 -07 6G -24 83* -21 32 -05 50 44 -Ol 25 17 -43 82* -44 56 00 32 31 -18 27 36 -35 -- -24 17 -28 20 34 -37 45 SB -06 81* 12 -11 -37 -10 14 -64* 29 19 -33 75* -30 10 -21 22 41 -27 12 26 06 -36 -12 -19 3O 04 44 52 25 34 29 -75* 44 -35 -09 -14 -23 18 -13 6D 50 -37 85* —48 -41 03 -04 O7 -20 35 70* -24 -- -35 -30 -Ol 16 -19 -02 3 42 O9 57 -20 -32 37 04 -12 54 ZB -17 49 -23 61 05 23 27 -11 -;; 8+9 -13 18 -12 79* OO 42 -O6 01 -26 24 -50 17 -35 -- -24 22 -39 -16 -28 13 25 00 -03 -61 38 25 08 37 12 SD 58 -33 00 -50 83* -33 21 07 00 20 39 -28 -30 -24 -- -16 41 31 O9 7 43 -44 -02 -02 67* 38 23 57 21 1 47 -34 15 -43 43 -83* 11 -39 -21 32 -24 09 12 43 -30 75* 14 44 11 383 -23 20 -01 22 -16 -- -03 64* 05 25 -01 -04 15 -13 10 77* 25 72* 18 14 -31 34 02 -O9 -37 74* -14 48 -09 29 22 29 16 -63 O4 43 38 39 35 6A -03 34 -23 -37 33 27 37 29 53 2A 04 -55 18 O3 -36 -22 -81* -14 -32 QB 41 34 16 -39 41 -03 -- 00 59 10 67* -22 34 -59 46 03 72* 28 60 6E -25 -24 -35 12 4O 62 -15 81* -33 15 -14 -37 -19 -15 31 64* 00 -- -18 6B 19 45 -02 -28 09 05 59 -18 -- 2C -16 25 -21 07 00 03 38 -1O 80* ** All decimal points omitted. 49 Table 9 (cont'd) Vars. RF RF RF RF RF ——x -—xi -—xii ——xiii ——xiv 16 35 ~28 25 ~53 27 18 35 ~01 ~06 27 22 21 27 32 ~34 45 29 6C 08 ~13 24 ~01 ~12 6G ~30 05 08 00 ~03 17 ~34 26 07 22 ~37 36 ~28 21 05 22 ~15 5B 09 36 22 31 ~15 19 10 04 07 08 11 26 ~26 ~36 ~49 ~09 33 34 42 ~42 16 ~28 17 6D 17 ~36 37 ~49 36 35 25 ~16 53 ~22 03 3 -07 ~51 52 ~24 ~21 2B ~41 10 ~08 02 07 8+9 ~31 ~17 23 ~24 ~09 24 ~53 ~01 29 ~20 ~46 13 61 ~20 ~15 ~20 46 5D 69* 37 ~38 58 15 20 34 31 ~59 55 13 7 02 ~37 ~14 ~05 02 1 68* 54 ~42 64* 28 32 ~47 ~52 49 ~63 ~27 38B ~59 ~60 48 ~70* ~07 25 08 ~34 43 ~43 ~25 14 ~17 ~35 39 ~65* 12 29 ~01 ~29 ~09 ~20 44 6A ~02 21 ~10 25 ~30 2A ~06 ~24 21 ~31 ~07 QB 04 19 ~34 46 15 10 01 ~24 ~32 18 24 6E 30 ~21 12 ~36 00 15 04 ~45 ~04 ~54 24 63 ~25 ~15 ~22 38 ~12 2C ~37 ~13 ~12 26 ~41 Vars. 30 30 38A 5A 23 28 33 27 12 5C 22 6F 11 31 RF. -—1 35 ~24 ~22 29 ~01 ~06 20 ~33 ~02 22 05 09 22 ~21 RF —X -79* ~59 42 29 07 24 ~15 ~25 25 ~07 ~22 15 ~11 RF.. —11 ~28 61 31 22 21 05 ~38 ~05 41 22 ~51 28 ~15 ~15 29 ~20 31 78* ~18 ~43 00 ~30 74* ~34 69* ~36 43 * p